Corning Credit Union adds plain-language search with Tursio

Corning Credit Union has connected its core banking data to Tursio, giving staff direct, plain-language search access to member data.

The change replaces a process that often took two to three weeks for non-urgent data requests. Those requests previously went through an intake committee before being scheduled into a sprint or placed in a backlog.

Tom Foster, Vice President of Product, Project Management and Data Analytics at Corning Credit Union, said the delays often reduced the value of the information by the time it arrived.

"By the time the data could be delivered, the requester had often lost focus on the problem," said Foster. "A lot of questions simply didn't get asked because it wasn't worth the trouble."

The deployment centres on the credit union's core system, Jack Henry's Symitar, which stores records by account rather than by person. As a result, a single member can appear under several name records, with small differences in addresses or other details.

Corning had already built an internal householding process to address that issue during a Salesforce rollout. The process groups accounts under the member they belong to. Foster said that work also proved important for the search platform because it made it possible to answer questions at the member level rather than through a series of account joins.

Existing systems

Instead of creating a separate data pipeline, Corning connected Tursio to data cubes drawn from its on-premises warehouse and Salesforce Data360 environment. Those cubes combine Symitar records with information from digital banking and other internal systems.

Foster said the credit union already used those same data sources for reporting through SSRS, Tableau and Salesforce. Using that consolidated source for the new search system was meant to keep the underlying data consistent across reporting and operational use.

Early users include product managers, who are using the system to size member segments for marketing offers. Support staff have also begun using it for operational checks, including identifying members approaching their 18th birthday or finding account combinations that should not exist.

Branch leadership is expected to be the next group to receive access, allowing managers to identify member relationships that need attention without submitting formal requests to a data team.

Rollout pace

Foster described the implementation as deliberate, with the credit union focusing first on data trust and user understanding before broadening access.

"We only have one chance for Tursio to make a first impression with the team," said Foster. "Success with Tursio depends on the team trusting the data it provides."

Before extending the rollout, Corning spent time training the model on what it knows and does not know, as well as on the appropriate use of its answers. Foster said that preparation was designed to reduce the risk of incorrect assumptions about the system's output.

Data privacy was a central factor in the search for a tool of this type, Foster said.

"In early 2025, we wanted a solution that completely walled off our data from anything external," said Foster. "Tursio was the most compelling option for data privacy."

Tursio said the credit union's approach reflects a wider issue for financial institutions considering artificial intelligence tools for member information. Concerns about exposing sensitive data remain one of the main barriers to adoption.

Alekh Jindal, Chief Executive Officer and Co-Founder of Tursio, said the product is designed to operate within a customer's own security boundary.

"The barrier we hear about constantly is the risk of breaching sensitive member data," said Jindal. "Our approach is to bring the AI to the data instead, so it runs inside the credit union's own security boundary and nothing leaves that environment. Corning had already done the hard work of organizing their data around members instead of accounts, which made that connection straightforward."

Corning currently uses the system for ad hoc analysis and data pulls. Foster said fraud detection is among the next planned operational use cases, while the credit union also intends to add trending member data and transaction-level data over time.

He said other credit union leaders considering a similar project should set clear expectations with staff before broad use begins.

"One misunderstood question can sour a user from returning and build on impressions about AI that are already in place," said Foster. "Be clear about what it can and cannot do, and what they should and should not do with the answers."

Corning Credit Union has more than 160,000 members worldwide and USD $2.4 billion in assets.